R 编程：从 data.frame 中选择每个因素组合

Question

我有一个 data.table 看起来像：

     Position              Name Salary
 1:       WR       Julio Jones   9300
 2:       WR Odell Beckham Jr.   9200
 3:       WR  Demaryius Thomas   9100
 4:       WR        Dez Bryant   8700
 5:       QB     Aaron Rodgers   8600
  ---                                  
 904:       TE       Jean Sifrin   2500
 905:       TE         Khari Lee   2500
 906:       TE       John Peters   2500
 907:      DST            Bears    2400
 908:      DST          Raiders    2300

我想得到所有可能的球队组合，由 1 个 QB、3 个 WR、2 个 RB 和 1 个 TE 组成，其中没有重复的球员。我不知道如何在 R 中选择组合，任何方向都很好。

我从这里得到了包含数据的 csv 文件：https://www.draftkings.com/contest/draftteam/7962690

如果你下载它并想玩数据，这是我目前的代码...

library(data.table)
library(dplyr)
dk <- read.csv(".../Downloads/DKSalaries.csv")
dk.dt <- as.data.table(dk)
dk.dt <- select(dk.dt, Position, Name, Salary, AvgPointsPerGame)

最佳

Answer 1

使用较小的数据样本，这里有一些代码（不太有效！）应该提取这些分组。

## Create a smaller sample to work with, 4 of each position
dk <- read.csv("DKsalaries.csv")
dk <- lapply(split(dk, dk$Position), function(x) x[sample(4), ])
dk <- dk[-1]                         # remove the other position
dk <- dk[c("QB", "WR", "RB", "TE")]  # reordering, for no reason really

## Expected number of combinations
## #QBs * choose(#WR, 3) * choose(#RB, 2) * #TE
4*choose(4,3)*choose(4,2)*4
# 384

## Get indices of combos within each group
rows <- list(t(1:4), combn(4,3), combn(4,2), t(1:4))  # these are possible combinations of each position
dims <- sapply(rows, NCOL)
inds <- expand.grid(mapply(`:`, 1, dims))             # indicies of combinations in 'rows'
dim(inds)
# [1] 384   4

## Function to extract a group
extract <- function(ind) {
    g <- inds[ind,]
    do.call(rbind, lapply(1:4, function(i) dk[[i]][rows[[i]][,g[[i]]], ]))
}

## So, one combination would be 
extract(1)
#        Position              Name Salary           GameInfo AvgPointsPerGame
# QB.5         QB     Aaron Rodgers   8600  GB@Chi 01:00PM ET           23.428
# WR.1         WR       Julio Jones   9300 Phi@Atl 07:10PM ET           21.293
# WR.3         WR  Demaryius Thomas   9100 Bal@Den 04:25PM ET           22.812
# WR.2         WR Odell Beckham Jr.   9200 NYG@Dal 08:30PM ET           26.417
# RB.13        RB    Jamaal Charles   7900  KC@Hou 01:00PM ET           17.093
# RB.20        RB      Arian Foster   7600  KC@Hou 01:00PM ET           22.808
# TE.191       TE      Travis Kelce   4800  KC@Hou 01:00PM ET           11.825

然后，要获得列表中的所有组合，您可以这样做

res <- lapply(1:384, extract)

R 编程：从 data.frame 中选择每个因素组合

R programming: choose every combination of factors from a data.frame

r

combinations

factors